TPS62110 locking up

The power profile attachment did not come through for some reason.  Could you try re-attaching it?

You don't need to toggle the SYNC pin for high currents.  You can leave it low and the IC will transition out of low power mode on its own.

You say transitioning SYNC from low to high makes the part work again?  Can you show a scope capture of Vin, Vout, SW, and SYNC when this happens?

Can you send me your e-mail? I was told not to post the power profile to general public to see.

We are toggling to constant frequency to reduce Vcc noise not because of higher current consumption  (peak current on this regulator is about 100 mA@ 7Vin 3.3Vout)  

Low power setting was introducing noise to an RF transmitter so changing to Constant frequency eliminated the ripple just on TX radio time when  needed.

Hi power is enabled about 5 to 20 times each second in 15ms windows , so duty cycle is quite low.

Yes, transitioning Sync from low to Hi makes the part work again (we just cut the control line to sync and let the pull up resistor get sync signal HI )

Next time another Unit fails will, take scope capture of all signals.

Problem  is that is a random event, and the Units took weeks to fail, unfortunately with the customers in other states.

You can request that I be added as a friend.  Then, we can start a 'conversation' which is private.

Next time it fails, just capture a waveform of Vin, Vout, SW and SYNC.  Then, capture a waveform of those same signals when SYNC is taken high again.  Finally, a waveform of normal operation with SYNC toggling would be good to see now so that I have a clear picture of what you are doing.

As a workaround, you can just keep SYNC tied high.  There's also our newer TPS62160 and TPS62170 which have much lower noise power save modes and come with much smaller solution sizes.

Tomorrow will send you the scope signals for the normal operation.

Did Keeping SYNC tied high worked  for the guy on the following thread?

http://e2e.ti.com/support/power_management/non-isolated_dcdc/f/196/t/100762.aspx

We will like to have some assurances to implement the workaround  or to do something more drastic.

Thanks

Yes, not toggling the SYNC pin was a solution for that customer.  But we need to make sure you have the same issue.  So, the waveforms would be great to see.

Hi

Here are the signals on normal operation of our product.

CH1 Vin

CH2 Vout

CH3 SW

CH4 SYNC

Thanks

Hi 

We just got another unit  on the lock up state

I wired up to take the signals but unfortunately this one did not wake up when the SYNC line was toggled to Vcc(6V) 

We are keeping the unit energized

What should be the next step to debug this?

Thanks

Jorge 

Did you capture a waveform of the same signals when it went into the state?

Does driving the SYNC pin from low to high reset the device?  If it did before, what is different about this setup?

we cannot capture the waveform because its random, this unit took 3 weeks to fail and had to bring back from the field

The setup is exactly the same, I was expecting the unit to recover when I wired a cable from VIN to SYNC, but did not.

Difference from last time is that this time we wired VIN to SYNC instead of removing R194 and letting the 150K pull up to take SYNC high

You'll need to have the probes in place while its operating and trigger on Vout or the SW node to capture the waveform of it failing.

Can you try removing R194 now?  How did you remove it in the past?  Did you de-solder it while the device was powered or something else?

Chris, do you want Jorge's phone number?

We have equipment dying in the field, I'm looking to ECO this away.  Current solution seems to be to stop using the low power mode, so if there's any useful information you can gather from this unit lets help while we can.

Mick (VPE)

Hi

The units that are failing are with the customer, we cannot wire them to see when they fail, we can wire units in our lab and Hope they fail

Removed R194, SYNC pin is now stable hi (same as Vin), unfortunately this did not wake up the switcher.

As I did not see anything else to do, I power cycle the unit and came back alive, find attached the wakeup sequence captured

Since this state seems to be entered by toggling the SYNC pin continuously for long periods of time, keeping the SYNC pin in one or the other state eliminates this issue.  An alternative to toggling the SYNC pin would be to get the SYNC pin in power save mode and instead toggle an external FET + resistor to add a load to the output that is high enough to put the IC in PWM mode and get the lower noise.

Cycling the EN pin should reset the part if this state is entered.  You might try that if you get it to enter this state again.

Toggling the SYNC pin should not reset the device, so it's curious that you saw it do this before.